| 1 | #!/usr/bin/env python3 |
| 2 | import hashlib |
| 3 | import json |
| 4 | import os |
| 5 | from pathlib import Path |
| 6 | import stat |
| 7 | import sys |
| 8 | |
| 9 | |
| 10 | root = Path(sys.argv[1]).resolve(strict=True) |
| 11 | if not root.is_dir(): |
| 12 | raise ValueError("Expected a directory") |
| 13 | digest = hashlib.sha256() |
| 14 | files = 0 |
| 15 | size = 0 |
| 16 | |
| 17 | |
| 18 | def fail(error): |
| 19 | raise error |
| 20 | |
| 21 | |
| 22 | for directory, children, names in os.walk(root, followlinks=False, onerror=fail): |
| 23 | children.sort() |
| 24 | for name in sorted([*children, *names]): |
| 25 | entry = Path(directory) / name |
| 26 | relative = os.fsencode(entry.relative_to(root)) |
| 27 | mode = entry.lstat().st_mode |
| 28 | if stat.S_ISDIR(mode): |
| 29 | kind = b"d" |
| 30 | elif stat.S_ISLNK(mode): |
| 31 | kind = b"l" |
| 32 | elif stat.S_ISREG(mode): |
| 33 | kind = b"f" |
| 34 | else: |
| 35 | raise ValueError(f"Unsupported file type: {entry}") |
| 36 | digest.update(kind + len(relative).to_bytes(8, "big") + relative) |
| 37 | if kind == b"l": |
| 38 | target = os.fsencode(os.readlink(entry)) |
| 39 | digest.update(len(target).to_bytes(8, "big") + target) |
| 40 | elif kind == b"f": |
| 41 | files += 1 |
| 42 | length = entry.stat().st_size |
| 43 | size += length |
| 44 | digest.update(length.to_bytes(8, "big")) |
| 45 | with entry.open("rb") as source: |
| 46 | for chunk in iter(lambda: source.read(1024 * 1024), b""): |
| 47 | digest.update(chunk) |
| 48 | print(json.dumps({"sha256": digest.hexdigest(), "files": files, "bytes": size})) |